New Smarter Regulatory Sandbox developed to increase compliance
HSE and the Safetytech Accelerator developed the Smarter Regulatory Sandbox, a collaborative environment to test digital products and increase compliance.
Two construction workers in a building site
Background
Construction workers face a wide range of challenges – from physical health risks to exposure to hazardous substances.
Importantly, a range of factors, such as the diverse nature of construction projects, varying levels of expertise among contractors and complex supply chains, create fragmented health and safety practices, presenting unique challenges.
For example, smaller companies or subcontractors may lack the resources or knowledge to implement robust safety measures, leading to higher risk of accidents, injuries, and even fatalities. Therefore, protecting construction workers often requires going beyond regulatory compliance; it requires a proactive approach to creating safer, healthier work environments.
The Knowledge Asset solution
The team at the Health and Safety Executive (HSE) explored approaches to strengthen health and safety practices in the construction industry in order to reduce injuries, improve regulatory compliance, and foster a culture where workers’ wellbeing is prioritised at every level.
One of the findings resulted in a partnership with the Safetytech Accelerator to develop the Smarter Regulatory Sandbox (SRS), the first of its kind.
The groundbreaking Smarter Regulatory Sandbox applied digitised health and safety data to several health and safety challenges, increasing regulatory compliance, boosting efficiency across the construction sector and enabling the development of innovative products.
By combining a Sandbox approach with regulatory data, HSE and the Safetytech Accelerator created a flexible and collaborative environment where regulators, construction companies, and technology developers could come together to explore the potential of digital innovation, such as AI and robotics.
One example of this innovative approach involved exploring how using CCTV footage from construction sites not only to monitor project progress but also - when combined with regulatory data- to distinguish between compliant sites and those posing health risks to workers.
Who will this help?
-
The construction industry: Making health and safety regulations machine-readable helps reduce the regulatory burden and provides clearer guidance on compliance.
-
The workforce: Better understanding of regulatory compliance and application of safety standards enhances protection from work-related injuries
-
Tech companies: Combining the Sandbox approach with regulatory data, supports innovation, creates a collaborative space to test digital products, and understand their real-world impact.
-
Regulators: Creating a safe and collaborative environment allows exploration of how new technologies affect compliance and safety outcomes.
Funding awards
The HSE was awarded £249,580 in the ‘Extend’ band of the Knowledge Assets Grant Fund in September 2023.
GOTT’s role
GOTT provided grant funding to the project.
Early results
The SRS project provided several positive results:
-
Accessing content directly from the regulator improved the accuracy of the Large Language Model by 30%
-
Using CCTV images helped track construction progress against plans and identify potential risks in advance
-
Creating common data standards facilitated the development of automatic compliance checks
-
Using AI-driven compliance solutions positively impacted the construction sector, although accessing quality source data remained challenging
-
Trialling a workplace fatigue app helped detect acute fatigue, provided user feedback to improve performance, and highlighted gaps in fatigue management
Next steps
The team continue to make products enhancements and collaborate with industry partners.
The next steps for the SRS project include:
-
Continuing collaboration with industry partners to broaden health and safety data and improve AI models for better compliance checks
-
Adding synthetic knowledge to increase prediction accuracy and help AI better understand compliance data
-
Improving structured audit, assurance data, and training AI with more diverse datasets
-
Combining AI insights with human expertise to make compliance monitoring more effective and support the shift to automated checks
-
Developing a blueprint for a regulatory Sandbox design, so that other Regulators can apply learnings and insights gained through this work